Extracting Reputation with Knock-Out Tournament-Based Pairwise Elicitation in Complex Social Networks

Complex social networks are typically used in order to represent and structure social relationships that do not follow a predictable pattern of behaviour. Due to their openness and dynamics, participants in these networks have to constantly deal with uncertainty prior to commencing in any type of interaction. Reputation appears as a key concept helping users to mitigate such uncertainty. However, most of the reputation mechanisms proposed in the literature suffer from problems such as the subjectivity in the opinions and their inaccurate aggregation. With these problems in mind, this paper presents a decentralized reputation mechanism based on the concepts of pairwise elicitation processes and knock-out tournaments. The main objective of this mechanism is to build reputation rankings from qualitative opinions, so getting rid of the subjectivity problems associated with the aggregation of quantitative opinions. The proposed approach is evaluated in a real environment by using a dataset extracted from MovieLens. © 2013 Springer-Verlag.